Scope and goals

[bogovicj]

The framework in this library should:

1. Write data using the n5 api
   a. including multiscale pyramids
   b. be flexible with respect to how data are downsampled.
   c. facilitate parallel computation locally and with spark
   d. writing N images at once (parallelize over all blocks of N images)
2. Write metadata using the n5 api
   a. to a number of select metadata formats (e.g. bdv-xml, n5-viewer, ome-zarr, ...)
   b. Neuroglancer support